Schizophrenia is a psychotic spectrum disorder, which impacts multiple domains, including cognitive functioning, interpersonal relationships, impairments in daily activities and ultimately reduces the quality of life for affected individuals. As a chronic mental health condition, schizophrenia affects millions of people worldwide and leads to cognitive dysfunctions and abnormal behaviors in patients. In recent decades, Artificial Intelligence (AI) has revolutionized healthcare, by making remarkable contributions towards disease diagnosis, personalized treatment planning, and enhanced patient care outcomes. Therefore, we presented an AI framework using Machine Learning (ML) and Ensemble Learning (EL) models along with Feature Selection to predict prodromal symptoms in Schizophrenia patients. We utilized an Open-source dataset collected from 5000 patients comprising clinical, psychological and behavioral symptoms. Among all the developed classifiers, the customized EL-based STACK models achieved best results for Accuracy, Precision, Recall, F1-score and Average AUC of 96.2%, 96%, 96%, 96.7%, and 93% respectively. Recently, Explainable AI (XAI) techniques are gaining attention in making classifier predictions more interpretable, understandable and reliable. Therefore, we employed XAI-based Shapley Additive exPlanations (SHAP) architecture, for generating visualizations including Violin, Waterfall, Force and Dependence Plots, which deliver meaningful interpretations of proposed classifier predictions. The motivation of this research study is to correctly predict the prodromal symptoms of Schizophrenia disease by employing EL-based stacking classifiers integrated with XAI tools, which can help the clinicians in making informed decisions.
Roopalakshmi et al. (Sat,) studied this question.